Reputation: 138
I have a "rel_type" table, with 3 fields: id, id_customer, id_type, I must uniquely get the id_customer that have only values 9 and 11 as id_type. So if I have recods that have these two values but have others, they don't have to be extracted.
I have tried in various ways but without results.
For example:
id | id_customer | id_type
--------------------------
1 | 123 | 11
2 | 345 | 9
3 | 123 | 9
4 | 788 | 5
5 | 788 | 11
6 | 788 | 9
7 | 788 | 4
I expect this output: 123, which is the only id_customer that has the two requested values and no more than these.
Upvotes: 0
Views: 551
Reputation: 164089
You can group by id_customer
:
select id_customer
from tablename
group by id_customer
having min(id_type) = 9 and max(id_type) = 11 and count(distinct id_type) = 2
Upvotes: 0
Reputation: 311188
I'd group by the id_customer
, and then apply two conditions - that the total count is 2 and that the count of records with 9 or 11 is also 2:
SELECT id_customer
FROM mytable
GROUP BY id_customer
HAVING COUNT(*) = 2 AND
COUNT(CASE WHEN id_type IN (9, 11) THEN 1 END) = 2
Upvotes: 1